Knowledge Reduction And Rule Acquisition

Classical rough set theory focus on discrete decision table, in which the valueof condition attributes is just a symbol that describe a feature of one object. Ex-tended researches on it have been done and some results have been made presented.However, there are also many continuous decision tables, which is di?cult to dealwith or get generalization knowledge by classical rough set models.In this paper, we are concerned with applications of extension rough set mod-els in continuous decision tables, especially attribute reduction and decision rulesacquisition in continuous decision tables. The paper is organized as follows:In chapter 3, a knowledge reduction method based on fuzzy clustering is studied.In the fist place, a method of computing the classification quality is given. secondly,an attribute reduction algorithm based on the classification quality are proposed.Then the reason to use support vector machine to mine decision rules is presented.Finally ,we use an example to show the process of these algorithms.In chapter 4, some concepts and properties of linear separability and convexhull are given at first. Secondly, a new rough set model based on convex hull isproposed. Then, an algorithm for computing the classification quality is presented.Next, an attribute reduction algorithm is put forward according to the classificationquality. Finally, an example is showed to explain the process of rule acquisition.In chapter 5, the definition of the covering similarity and the definition of thecovering set are given at first. Secondly, a new rough set model is proposed. Then,an algorithm for computing the classification quality is presented. Next, an attributereduction algorithm is put forward according to the classification quality. Finally,an example is taken to illustrate the process of rule mining.In chapter 6, the vector method of attribute reduction is presented.